Official Community BETA of the EQEmu Server Bots Begins Today!
 

I am pleased to announce that the community beta of the EQEmu Server Bots is ready for QA testing and begins today. This means that this code base that is going to replace what has been refereed to as “EQOffline” in about 7 days and before this happens, this beta period is the communities opportunity to try out this code and post issues that need to be addressed. Even if you do not use bots at all, you will want to download this code and try it as a lot of code was re-factored that could possibly impact that standard code, though I know of no issues right now nor do I expect any.

You can read more about the beta period and what to expect, here: http://wildcardx.blogspot.com/2009/0...mu-server.html

This is from the current Build


In tutorialb not grouped with bots I con the guards and they con amiably, I group up with bots they still con amiably but now they will attack on site, this only started to happen from today's commits. I have tested 3 times same results all 3 times, as long as I'm not group I can walk right up to them. :shock:

Can you provide any additional details? I tried to repeat this on my server but the guards don't attack me no matter if i am grouped with bots or not.

I'm really not sure what i can add to this......
EQEmu R850
PEQ DB 781
NON-GM account

Toon is level 8 I walk between the first 2 guards and nothing as normal, I group with a bot and do the same thing I'm Dead man walking...this happens every single time. I just tested with all quest givers and guards in Tutorialb and if you have a bot in group you are dead.

Maybe the name/NPCID of the guards, and the race of your character and the race of your bots would help? Though, I don't think any guards are supposed to be KoS to any races in Tutorialb.

Well I believe the Factions have nothing to do with it since they con amiably before I group and after I group, The con never changes just their attitudes :)

ALL Guards in tutb
Human Bot Cleric
Human pally ME

Well trevius was on to something IT was my HE pally that was killed by every guard and My Human pally is just fine. :confused:

Does this happen with any guards in any of the other starting cities, like Qeynos, Felwithea, etc?

It appears to be only tutorialb :( I have checked a handful of the others and its just fine.

Hmmm.. this sounds like an awfully uber specific bug :-)

I'll try to recreate it on my dev server, though just to be diligent.

Is anyone able to reproduce this bug? I can't. I made a human male pally with a human female cleric and neither Guard Maddoc or Guard Hobart would attack me, grouped with the bot or not. Btw, both con indifferent to me on my dev server (using the latest PEQ world database as of today July 30). Are there different guards I should try?

It was a High Elf Pally That had the problem and the Human pally was fine, sorry for the wrong info

I made a high elf male pally with a human cleric bot and no guards attacked me in or out of a croup with my bot. Until someone else can replicate this bug I am going to assume it is some sort of config error or something.

WildcardX,

I have found a few rather nasty bugs.

When raiding Crushbone for a dragoon dirk, my bot group encountered the Fabled Ambassador DVinn. The fight of course happened on the spot where DVinn spawns. In order to position my cleric, druid, and enchanter bots to give them better LOS to DVinn I used #bot group summon. The first bug I encountered is that immediately upon the enchanter being summoned to my position, his pet immediately went crazy and basically killed DVinn in a matter of a couple of seconds. It performed about 50 melee rounds uninterrupted as though it were super-hasted.

The second bug is that the kill yielded no EXP and DVinn poofed.

The third bug is that #bot group summon appears to be able to instantly crash the entire zone. It does not always happen, but it is often.

Details:

Latest build from SVN as of this posting with latest PEQ database as of this posting (modded with Bots.sql from the same SVN update).

Thanks for all your hard work, and keep it up!

I haven't been able to recreate any of these issues. Can anyone else replicate these bugs?

WildcardX,

The bug affecting bots with pets seems to be related to pathing. I cannot replicate the bug UNTIL the bot's pet ends up against a wall or other object. The little hut in Crushbone with the Orc Scoutsman is where I can get it to happen with great frequency. When the pet gets "hung" on the back wall, the endless attacks occur. Also it seems to happen most when the pet is positioned behind the mob.

Also I have an issue with 'bots' table in the database. I cannot remove bots from the table (via Empty or Truncate with NaviCat) due to foreign key constraints. This may be related to some ineptitude on my part, but I cannot get it to work.

Quick question: Do you plan to add GM bot editing commands since #npcedit will no longer apply to bots?

I think the work you are doing is great, WildcardX. Please keep it up!:D

More testing...
Ok this time I used a Barb BL and for the bots I had Barb warrior, and DE cleric. I have screen shots to maybe help understand what I'm talking about. Everything is a fresh install, Eqemu rev 880 and PEQ Rev859.
http://lamarch.org/downloads/bot3.jpghttp://lamarch.org/downloads/bot1.jpg

WildcardX,

I've found a few more things that might need looking at.

Using '#bot spawnraid' can spawn the bots, but have them appear 'grayed-out' in the party window (as though they were clients in another zone).

There seems to be an aggro problem (a HUGE one) possibly related to leslamarch's problem. Take your bot group to Mistmoore and attack something. Soon you will have what appears to be the entire zone upon you. I couldn't confirm it was the whole zone, but it looked like what happens with #aggrozone.

Giving a bot a weapon with a proc that they cannot proc yet yields the message "Your pet's will is insufficient...".

There is a bug in that the client sends something like OP_FloatListThing (approximate name...not sure on that) and that results in the client getting disconnected.

Cancelling trade with a bot can make an item appear to *poof* (not visible in the client inventory or on the cursor). When you attempt to trade another item to the bot, it gets the item that poofed.

Bots appear to be running past the mob, in a straight line in the direction they are facing, but then warp back to their actual location (looks like classic lag effects between clients).

One of my human monk bots has a different face each time she shows up. lol

Nice screenshots :-) I don't know why I can't get this to happen for me, though. If I could get it t work then i can debug it and fix the issue. I am sure its probably something small and overlooked.

As for your issues, Taurinus, the constraints in the tables are to enforce data intgerity. You'll find that if you first delete the bots records in other bot tables, like botinventory, then you can go ahead and delete your bots.

At this time, I do not plan on implementing any time of #botedit commands as I don't believe bots are going to be editable in the same sense you would edit a NPC.

I think the issue about the weapon not procing because the bot isnt sufficient level to use the proc is by design and not an issue.

I am aware of the issue with cancelling trade. its annoying that it happens but I am planning to overhaul how bots calculate their stats and their inventory before I get to an issue like this one. In the meantime, I guess just work around it :-)

As for bots running past their target, what rev where you using when this occurred. I don't believe this happens in r880, but you tell me.

And your right, buts dont maintain their appearance at all. Again, its annoying, but I will get to fixing it after I merge back to trunk.

The bots code is far from perfect. There are many known issues, especially regarding how bots calculate their stats and how they access their inventory. I do plan to address all these issues and get them right, but I have to prioritize as I am only one man and I do have a day job I have to write code for. If I get some help with developing then maybe I can dibby up the work and we can perfect this system faster.

For now though, I am not aware of any critical or severe issues with the bots code. While it isn't perfect, neither was the system this is replacing. As i do not see any evidence of any critical issues like zone crashes or anything that is negatively impacting the normal, non-bots server logic, I am going to go ahead and merge this system to trunk and remove the EQBOTS code.

I believe I have this bug identified and fixed in r881. Go ahead and give this a try now.

WildcardX,

Yes, that fixed the #bot groupraid issue.

The issue about the bots and the procs is that I didn't think that the client should be getting messages about pets in reference to bots...but then again that is what they are afterall I suppose.:grin:

I really appreciate and enjoy the changes you've made to the bot behavior especially the rogue. He's even nimble enough to path behind and stab an npc while soloing it!

I do have coding experience (about 7 years of it) with c/c++ and intel assembly. I might dig in and see what I can do.

Thanks for the quick response.

Thats cool to hear your a developer and thanks for the praises! If you come up with any code for it, please send me the diff so I can see what you can do. I can definitely use assistance with bots. There is so much to be done but I just don't have the time to do it all myself. PM me if you have code to share for improvements and I'll provide you with my email you can send it to and we can discuss.

I'm just about to start using this new code update. I do ALOT with Bots in many different zones I'll see what I can come up with.

Laxa

oh another quick question,

I'm reading through the changelog, does this mean I need to recreate all my bots?

I need a little help with Linux compile please, I downloaded EQEmuServerBOTS and I run make from that dir and everything works except Bots, I used the Linux guide made by Hunter and I have Debian running. Any help would be much appreciated!

I think for linux compiles, you need to edit your makefile.common in the zone project and remove the "petAI.o" and "botAI.o" and then add "bot.o". Then in your makefile for both zone and world, you need to replace "-DEQBOTS" with "-DBOTS". Let me know if this works.

Thanks a million, I'll try as soon as I get off work today!!

Well, No Bots on my Linux yet, lol. I've added bot.0 to the makefile.common but petAI.o and botAI.o where not there to delete. Then I searched for -DEQBOTS in makefile in zone dir and none found so I added -DBOTS, after compiling I had no zone, launcher or world file that it had created. So I changed -DBOTS to -DBOT and it compiled all files but no bots in game, the command is not recognized. Everything else in game works fine. Any suggestions?

OK, I updated to the latest revision last night (after the merge of the branch back into the trunk), and here is a bug report from my Dev, Kayen:
So far that is the only thing that needs quick attention that we have seen so far. I will keep an eye out for more issues, but we will probably have to revert back until this issue is resolved.

If any of you can help identify these bugs, send me code so I can look it over. Otherwise I'll get to them as I can.

Since I am at work, it is hard for me to be of much help. I just went ahead and reverted it since we have dozens of high end events that rely heavily on $combat_state. I might have to start trying to read through the diffs of your merge to see if I can find anything though.

I just committed a fix for the COMBAT_EVENT signalling bug. That issue should be resolved now as of r888.

r888 also fixes the bug that bugged when you cancel a trade with a bot, even though I forgot to list that bug fix in the log.